Hermet Park [Thu, 11 Jun 2020 07:10:36 +0000 (16:10 +0900)]
evas glengine: fix code sync after migration.
Introduced a wrong result after upstream migration,
since we have tizen only code diff....
Change-Id: Ia4bac99a244a63a1d54c63dd93aea80c5a0bfce0
Carsten Haitzler (Rasterman) [Wed, 10 Jun 2020 18:30:14 +0000 (19:30 +0100)]
efl - build again - test suite break in using undefined ERR macro
Stefan Schmidt [Tue, 9 Jun 2020 13:21:28 +0000 (15:21 +0200)]
tests: eina: make sure we always check on eina_matrixsparse_data_idx_get return
It has been checked in all other places, but has been forgotten in these
two.
CID: 1401081
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11960
Stefan Schmidt [Tue, 9 Jun 2020 13:09:08 +0000 (15:09 +0200)]
tests: eo: handle error return of eina_value_int_convert()
CID: 1400962
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11958
Stefan Schmidt [Tue, 9 Jun 2020 12:04:31 +0000 (14:04 +0200)]
tests: eina: also handle failure cases when cleaning up created files/folders
Somethign fishy is going on if we can not delete the files and folders
we created.
CID: 1400986
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11955
Stefan Schmidt [Wed, 10 Jun 2020 12:12:59 +0000 (14:12 +0200)]
ecore_drm: check return of drmModeSetCrtc()
Report if we fail to reset the Crtc during output free. Only print an
error and keep going with the rest of the output free.
CID: 1402668
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11965
Christopher Michael [Wed, 10 Jun 2020 12:38:02 +0000 (08:38 -0400)]
tests/ecore_wl2: Fix unchecked return value
Coverity reports that we do not check the return of eglInitialize
here.
Fixes Coverity CID1412362
Christopher Michael [Wed, 10 Jun 2020 12:32:40 +0000 (08:32 -0400)]
tests/ecore_wl2: Fix unchecked return value
Coverity reports eglChooseConfig is called without checking return
value here. This patch fixes the issue by checking the return value
and by also making _init_egl return a bool so that we can fail the
test if eglChooseConfig results in failure.
Fixes Coverity CID1412365
Christopher Michael [Wed, 10 Jun 2020 12:26:29 +0000 (08:26 -0400)]
ecore_wl2_tests: Fix formatting
Stefan Schmidt [Tue, 9 Jun 2020 13:17:36 +0000 (15:17 +0200)]
tests: eina: check return of eina_module_load and unload
These could fail as well. Check in testsuite.
CID: 1400961
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11959
Stefan Schmidt [Tue, 9 Jun 2020 12:29:35 +0000 (14:29 +0200)]
tests: evas: check fd being non -1 before using close()
eina_file_mkstemp would return -1 in an error case. Make sure we check
the return when clsoing here.
CID: 1400790
Reviewed-by: Christopher Michael <devilhorns@comcast.net>
Differential Revision: https://phab.enlightenment.org/D11957
Vincent Torri [Wed, 10 Jun 2020 05:07:47 +0000 (05:07 +0000)]
Use extern after EAPI
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11961
Hermet Park [Wed, 10 Jun 2020 11:30:19 +0000 (20:30 +0900)]
evas stack: ++safe code.
null check to prevent any crashes (just in case)
Change-Id: Ic77a2c785feba9be32eb1460439f4c3315087e07
Hermet Park [Wed, 10 Jun 2020 06:49:49 +0000 (15:49 +0900)]
ui textpath: updated doc.
Change-Id: I562734a64b5ad8f371c1952a996a80f54a05e0c7
Taehyub Kim [Wed, 10 Jun 2020 06:39:03 +0000 (15:39 +0900)]
efl_ui_textpath: add the efl_ui_textpath_text_user_style_set API for applying the user text style
Summary:
For the text style customizing, added the new API to apply user text style.
The following example is the user style sample.
DEFAULT='font_size=16 color=#F00 underline=on underline_color=#00ffff'";
{
F3899541}
Test Plan:
1. run elementary_text (textpath)
2. toggle user style check box
3. see the result
Reviewers: Hermet, kimcinoo, jsuya, bu5hm4n
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11953
Change-Id: Ia3a182aebadcad24383c3021e358d4e40e5e9515
Marcel Hollerbach [Tue, 9 Jun 2020 10:09:25 +0000 (12:09 +0200)]
mono: update window construction
they simply missed parameters here
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11954
Stefan Schmidt [Tue, 9 Jun 2020 07:11:52 +0000 (09:11 +0200)]
ci: travis: disable the webp loader on CI builds for now
With the bump of the needed libwebp version to 1.1.0 we have have our
Debian and Ubuntu builds failing with a to old version. Disable by
default in CI for now and check back later.
Carsten Haitzler (Rasterman) [Tue, 9 Jun 2020 09:13:04 +0000 (10:13 +0100)]
efreet - handle runtime relocation right with default XDG_DATA_DIRS
XDG_DATA_DIRS was only set up to a default including where efl was
installed prefix-wise as the compiled-=in prefix, not runtime
determined prefix. it shouldn't actually affect most people except
those making use of this.
@fix
Carsten Haitzler (Rasterman) [Sun, 7 Jun 2020 00:33:43 +0000 (01:33 +0100)]
eina - prefix - use smaller buffers for building strings
use less stack space but no features as buffers are big enough for all
content used or alloca'd now.
Carsten Haitzler (Rasterman) [Sun, 7 Jun 2020 00:28:18 +0000 (01:28 +0100)]
eina - abstratc content - smaller stack buffer for limited size string
no need for a 4k buffer when 128 bytes will be plenty - short string.
Marcel Hollerbach [Sat, 18 Apr 2020 17:43:11 +0000 (19:43 +0200)]
efl_ui_win: declare constructors not optional
they cannot be optional, or in later languages types will collide.
Differential Revision: https://phab.enlightenment.org/D11741
Felipe Magno de Almeida [Mon, 8 Jun 2020 21:05:49 +0000 (18:05 -0300)]
cxx: Fix uses of intrinsic eolian binbuf type
Add special binbuf and Eina_Strbuf conversions
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11950
Marcel Hollerbach [Sun, 19 Apr 2020 13:14:12 +0000 (15:14 +0200)]
eo-files: convert all Eina.Strbuf & Eina.Binbuf
these structs should not be used, but rather the builtin types that
exist for it.
Reviewed-by: Felipe Magno de Almeida <felipe@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11742
Felipe Magno de Almeida [Mon, 8 Jun 2020 21:32:22 +0000 (18:32 -0300)]
dotnet: Ignore Efl.Object.parent as constructor
In C# we already have Efl.Object.parent as an implicit constructor. Ignore it if it is marked as a constructor.
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11951
Marcel Hollerbach [Sun, 19 Apr 2020 19:34:17 +0000 (21:34 +0200)]
eo: make the parent a optional constructor
this is needed in order to sanely bind parent calls to some objejcts,
but not all.
However, from the canvas structure up we *need* a parent, so we can
fetch the evas from it. So declare it there a none optional
Reviewed-by: Felipe Magno de Almeida <felipe@expertisesolutions.com.br>
Differential Revision: https://phab.enlightenment.org/D11734
Stefan Schmidt [Tue, 9 Jun 2020 07:43:57 +0000 (09:43 +0200)]
Revert "ci: disable wepb loader until we have a change to update the mingw cross env"
This reverts commit
63b1334e2c98f0c0f08801ce77d66f372e7abb30.
No longer needed as we have updated the mingw image to have 1.1.0
available.
JunsuChoi [Tue, 9 Jun 2020 01:57:20 +0000 (10:57 +0900)]
evas_vg_cache: Don't caching vg file if value_provider is applied.
Summary:
value_provider can change the value of the property received from VG file.
When a file is cached, the changed properties are applied to all other objects using the same file.
So. If value provider is applied, evas_vg_cache is not caching vg file.
Test Plan: N/A
Reviewers: Hermet, herb, kimcinoo
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11924
Ali Alzyod [Tue, 9 Jun 2020 01:25:10 +0000 (10:25 +0900)]
evas_common_font: release reallocated glyphs bitmaps data
Reviewers: woohyun, smohanty
Reviewed By: woohyun
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Maniphest Tasks: T8743
Differential Revision: https://phab.enlightenment.org/D11945
Myoungwoon Roy, Kim [Tue, 9 Jun 2020 05:08:11 +0000 (14:08 +0900)]
doxygen docs: Fix API reference group to Ecore_Eo and eina_vector
Change-Id: Ic20831f0db719f935dbf4e7421c70f7c4cdbd5cb
Subhransu Mohanty [Mon, 8 Jun 2020 11:28:14 +0000 (20:28 +0900)]
elementary/example: Fix memory leak in gengrid test
Reviewers: Hermet, SanghyeonLee
Reviewed By: Hermet, SanghyeonLee
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11928
Subhransu Mohanty [Mon, 8 Jun 2020 11:27:42 +0000 (20:27 +0900)]
ecore_ipc: Fix memory leak
Summary:
As delete_me flag was not updated in the delete callback it dosen't
call ecore_ipc_client_del() on the client hence the leak.
Reviewers: Hermet, raster, bu5hm4n, CHAN
Reviewed By: Hermet, CHAN
Subscribers: CHAN, cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11927
Hermet Park [Mon, 8 Jun 2020 10:46:31 +0000 (19:46 +0900)]
canvas gl: remove out-dated compiler keywords
Stefan Schmidt [Thu, 4 Jun 2020 10:52:07 +0000 (12:52 +0200)]
exactness: record and print test time on success
Having the runtime of a test is interesting for debugging and
optimization. Printing them only on success for now.
Differential Revision: https://phab.enlightenment.org/D11935
Stefan Schmidt [Tue, 2 Jun 2020 14:03:00 +0000 (16:03 +0200)]
ci: travis: add more fold and timing markers in some forgotten places
In teh initial patch we forgot some interesting places we also want to
fold and get timing info on. Dependency installation, cross native build
step, coverity tools download and codecov run.
Differential Revision: https://phab.enlightenment.org/D11938
Stefan Schmidt [Thu, 4 Jun 2020 13:19:47 +0000 (15:19 +0200)]
ci: simplify ccache setup handling
We no longer need the special case for release ready. This was needed
for the way autotools setup directories, which we no longer use. (We are
not having a config.log anymore)
Differential Revision: https://phab.enlightenment.org/D11937
Stefan Schmidt [Mon, 8 Jun 2020 07:10:51 +0000 (09:10 +0200)]
ci: disable wepb loader until we have a change to update the mingw cross env
We now need webp 1.1.0 and our current mingw cross env only has 1.0.2.
Disable to keep CI working while we update.
WooHyun Jung [Mon, 8 Jun 2020 02:03:07 +0000 (11:03 +0900)]
webp: add version check in configuration time
Reviewers: Hermet, herb, bu5hm4n, stefan_schmidt
Reviewed By: stefan_schmidt
Subscribers: stefan_schmidt, cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11926
Hermet Park [Mon, 8 Jun 2020 02:00:26 +0000 (11:00 +0900)]
eldbus: fix the data size issue.
Summary:
dbus message requires the void* value for the message data,
the returned value of dbus_message_iter_get_basic() can be mismatched by to the architure,
This could bring the returned va_argues memory corrupted among the series of values.
We can use the defined type for exceptions.
Reviewers: herb, kimcinoo, jsuya, raster
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11940
Carsten Haitzler (Rasterman) [Tue, 2 Jun 2020 10:39:06 +0000 (11:39 +0100)]
evas gl - tune gl engine to use a bit less cpu overhead
we have used a fair bit of cpu to avoid gpu overhead and this tunes
the balance back a bit to throwing more at the gpu and less at the cpu
by reducing the number of pipes and max vertex counts per pipe by
default and only resetting pipe state vars if adding the first pipe
member.
@opt
Change-Id: Id102255c532ec7d116ca9a4d45f18ad69cde3829
Hermet Park [Mon, 8 Jun 2020 05:00:04 +0000 (14:00 +0900)]
ui scroller: fix broken scroller loop
image is streched since proxy fill/region size is mismatched
this is a regression bug introduced by
76b89b8e49f4b16ad490c04134a9bdddea27950f
Change-Id: I1eb4de61e6a5e080a2f350de30cb76d1db575429
Daniel Kolesa [Sat, 6 Jun 2020 21:29:23 +0000 (23:29 +0200)]
ci: drop lua from default binding set as elua is off by default
Daniel Kolesa [Sat, 6 Jun 2020 17:09:21 +0000 (19:09 +0200)]
build: disable elua by default, plus nicer detection
Elua is now disabled by default. There are some other changes:
1) Elua scripts are only installed if Elua is enabled
2) Lua bindings are only installed if Elua is enabled
3) Elua with interpreter is clearly experimental and will message
Stefan Schmidt [Fri, 24 Apr 2020 07:35:31 +0000 (09:35 +0200)]
emotion: remove all left overs from removed backends
We removed them in the last release. Make sure we cleanup all the
references and dead code still belonging to it.
Extra note on the Emotion_Module enum removal. While it was exposed in
the public Emotion.h header it is never used anywhere in the code and
has no purpose, thus it can safely be removed.
Reviewed-by: Carsten Haitzler (Rasterman) <raster@rasterman.com>
Differential Revision: https://phab.enlightenment.org/D11941
Marcel Hollerbach [Thu, 4 Jun 2020 08:23:29 +0000 (10:23 +0200)]
build: do not publish surface .eo files
they are privat, they should never be installed.
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11933
Marcel Hollerbach [Thu, 4 Jun 2020 08:03:44 +0000 (10:03 +0200)]
evas: correctly declare filter_internal as public API
I know this is internal. However it is public used in public .eo files.
Even if it is internal, and beta, we still need to expose the API in
order to have it correctly in the .so.
Reviewed-by: Stefan Schmidt <stefan@datenfreihafen.org>
Differential Revision: https://phab.enlightenment.org/D11932
Bowon Ryu [Fri, 5 Jun 2020 05:26:22 +0000 (14:26 +0900)]
edje: markup flag should be TRUE in legacy edje_object_part_text_set
Summary:
Through some APIs (elm_entry_editable_set, elm_entry_single_line_set, etc.)
located after edje_object_part_text_set(legacy)
If efl_ui_widget_theme_apply is called,
In edje_object_part_text_raw_generic_set, the legacy flag becomes FALSE.
And in this case, the logic works in the unintended direction
because the set_markup flag is FALSE.
Test Plan:
/*
gcc -o entry_example entry.c `pkg-config --cflags --libs elementary`
*/
#include <Elementary.h>
EAPI_MAIN int
elm_main(int argc EINA_UNUSED, char **argv EINA_UNUSED)
{
Evas_Object *win, *en;
elm_policy_set(ELM_POLICY_QUIT, ELM_POLICY_QUIT_LAST_WINDOW_CLOSED);
win = elm_win_util_standard_add("entry-example", "test");
elm_win_autodel_set(win, EINA_TRUE);
en = elm_entry_add(win);
evas_object_size_hint_weight_set(en, EVAS_HINT_EXPAND, EVAS_HINT_EXPAND);
elm_object_part_text_set(en, "elm.guide", "<font_size=32 color=#FFF>TEST</font_size>");
elm_entry_editable_set(en, EINA_FALSE);
evas_object_show(en);
elm_object_content_set(win, en);
evas_object_resize(win, 300, 200);
evas_object_show(win);
elm_run();
return 0;
}
ELM_MAIN()
Reviewers: woohyun, ali.alzyod
Subscribers: zmike, cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11868
Carsten Haitzler (Rasterman) [Thu, 4 Jun 2020 19:25:46 +0000 (20:25 +0100)]
evas - genetic update regions smart merge - merge v adjacent regions
regions sometimes produce box sets like:
// +---+
// | |
// +---+ +-------+
// | | | |
// +---+ +-------+---------+
// | |
// | |
// +-----------------+
so the upper-left 2 boxes can be merged into 1 and they have the same
x coords and are flush-aligned one above the other. that is what
this does - find these and merge them to have fewer rects
@fix
Carsten Haitzler (Rasterman) [Thu, 4 Jun 2020 19:23:13 +0000 (20:23 +0100)]
evas render - add debug to draw boxes around update regions
helps debug update regions to see if there is too much overdraw or too
many regions. i smelled too manhy and this showed it. i was right. fix
for too many regions coming next
Carsten Haitzler (Rasterman) [Thu, 4 Jun 2020 19:22:03 +0000 (20:22 +0100)]
tiler - add maxreg for newtiler - not used atm but matches current
Bartlomiej Grzelewski [Wed, 20 May 2020 09:12:38 +0000 (11:12 +0200)]
atspi: return proxy object
Proxy objects were never returned by function because
they do not have x,y,w,h coordinates.
Change-Id: I237e63b6da3bd345f74e92c9e135bb9ff8d17599
Taehyub Kim [Fri, 5 Jun 2020 08:14:13 +0000 (17:14 +0900)]
elm_poup_eo.legacy.h: modify Elm_Popup_Group to Elm_Popup
Change-Id: I4a49df80c7a92ba388683cb7f96ecbb6666fdf24
Carsten Haitzler (Rasterman) [Thu, 4 Jun 2020 11:00:52 +0000 (12:00 +0100)]
need to upgrade meson. it looks like older ones have trouble
like adding endless: -lm -ldl -lm -ldl ....
Carsten Haitzler (Rasterman) [Tue, 2 Jun 2020 10:37:37 +0000 (11:37 +0100)]
elm test - animated icon - big win to have a bg and not garbage
Marcel Hollerbach [Thu, 4 Jun 2020 01:29:44 +0000 (10:29 +0900)]
build: move static out of none_static
Summary: these are static libs, they shall not be in the static list.
Reviewers: stefan_schmidt, jsuya
Reviewed By: stefan_schmidt, jsuya
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11923
JunsuChoi [Thu, 4 Jun 2020 01:28:21 +0000 (10:28 +0900)]
Efl.Canvas.Vg.Object: Fix backup size
Summary:
Make the size to be backed up and the size to be compared equal.
and minor fix
Test Plan: N/A
Reviewers: Hermet, herb, kimcinoo
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11925
JunsuChoi [Thu, 4 Jun 2020 00:22:39 +0000 (09:22 +0900)]
Efl.Gfx.Path: Modify boundary calculation.
Summary:
The point of the path is of type float.
The boundary must contain the coordinates of the point,
so min_x,y value must be round down and the max value round up.
Test Plan: N/A
Reviewers: Hermet, herb, kimcinoo
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11917
Stefan Schmidt [Wed, 11 Dec 2019 14:51:19 +0000 (15:51 +0100)]
ci: log time in every fold start and end
Annotating the build log with timings will help us to understand better
where the time is spend during the builds and see areas to optimize.
Switching over to the "native" bash functions in Travis, so switching
all scripts to bash.
Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11913
Stefan Schmidt [Wed, 3 Jun 2020 08:37:38 +0000 (10:37 +0200)]
ci: remove no longer used address sanitizer leak suppression file
We switched the asan build to ignore leaks for the internal build tools.
Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11922
Stefan Schmidt [Wed, 3 Jun 2020 08:30:04 +0000 (10:30 +0200)]
ci: remove outdated README file
This is no longer up to date. Remove before it gives a wrong description
of things.
Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11921
Stefan Schmidt [Wed, 3 Jun 2020 08:28:12 +0000 (10:28 +0200)]
ci: remove not in use script
We are not buildign the external examples. This script is no longer
used. We have an tiny internal app build test and building the in-tree
examples as well.
Signed-off-by: Stefan Schmidt <s.schmidt@samsung.com>
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11920
Stefan Schmidt [Wed, 3 Jun 2020 07:18:02 +0000 (09:18 +0200)]
Revert "elementary: build elementary_test additionally as shared object"
We now longer need this workaround. Exactness is now pre-loaded and we
can run it with any executable without this trick. Remove to simplify
our build.
This reverts commit
78ad088dd73a18fb4cfcfedcc18a6bf03e28080b.
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11919
JunsuChoi [Wed, 3 Jun 2020 06:58:00 +0000 (15:58 +0900)]
Efl.Canvas.Vg.Object: Prevent null pointer access
add null check
Subhransu Mohanty [Wed, 3 Jun 2020 04:03:39 +0000 (13:03 +0900)]
evas/engine: Fix memory leak in generic cache.
Summary:
During shutdown we used to call engine_image_free() which was causing some deadlock.
as we have evas_cache which takes care of freeing all the images we just have to
delete the generic cache without freeing the image during shutdown.
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11915
JunsuChoi [Wed, 3 Jun 2020 02:37:23 +0000 (11:37 +0900)]
Efl.Canvas.Vg.Object: Optimize Ector Surface Size
Summary:
The ector surface size was determined by the size of the vg object.
vg object is usually sized by the size of the container.
So, the ector surface is set unnecessarily large.
This patch sets the ector surface size to the path boundary.
And the path boundary refers to the stroke width and miterlimit.
Test Plan:
vector sample
{
F3887634}
{
F3887632}
[grey area is ector surface size]
{
F3887633}
Reviewers: Hermet, kimcinoo, smohanty, herb
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11865
Subhransu Mohanty [Wed, 3 Jun 2020 02:17:18 +0000 (11:17 +0900)]
ecore: Fix Asan stack overflow warning
Summary:
As argument can be passed by register (depending on the compiler optimization)
when we take the adress of the nbytes and pass it to send() function which reades
4 bytes from it ASAN flags it as a stack overflow . So just assign the value to a
local variable to avoid the warning.
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11914
WooHyun Jung [Wed, 3 Jun 2020 09:53:43 +0000 (18:53 +0900)]
build: add gl_common files to engine_src
@tizen_fix
Change-Id: I06eb654aed79e19c3e945e586ea87dd44e1c0a78
Subhransu Mohanty [Wed, 3 Jun 2020 03:15:05 +0000 (12:15 +0900)]
elementary_test: Added --autoclose option to elementary_test
Summary:
This option will enable writing tools which will automate running the
elementary test cases.
Reviewers: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11910
Change-Id: I9c11fdaf44141e4610e63de363779021f382c251
Subhransu Mohanty [Wed, 3 Jun 2020 04:43:52 +0000 (13:43 +0900)]
elementary: Fix memory leak in elm_win
Change-Id: I1003d7e7abc31c0425e8c12553a20c77193cff01
Jongmin Lee [Wed, 3 Jun 2020 02:04:31 +0000 (11:04 +0900)]
elm_gengrid: resolve double free
Change-Id: I17789de33c088c6baeced2ea5758e6a34c37902c
Jongmin Lee [Tue, 2 Jun 2020 22:20:18 +0000 (07:20 +0900)]
elm_atspi: resolve migration build error
Change-Id: I175602b125bcfb1101011b37067b6b3cd1397576
Stefan Schmidt [Thu, 28 May 2020 17:32:03 +0000 (19:32 +0200)]
ci: travis: enable efl-one in our all enabled build
After a successful build we run the efl-one test script to see if it
drags in unwanted libs.
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11906
Marcel Hollerbach [Tue, 2 Jun 2020 06:28:27 +0000 (08:28 +0200)]
build: add eet to evas suite
evas suite seems to use internal headers, which includes eet.
Marcel Hollerbach [Tue, 2 Jun 2020 06:11:23 +0000 (08:11 +0200)]
ci: fix for rename luajit to lua bindings
this is an incredible hard fix. The rename was essential, and it was
important that the CI was killed for 3 days, the names in this script
was also hidden that deep that litterly *noone* could have ever renamed
them without a daylong hike through the masses of script here.
Marcel Hollerbach [Mon, 1 Jun 2020 20:32:53 +0000 (22:32 +0200)]
build: do not link test suite again against static libs
this might result in duplicated symbols.
Ali Alzyod [Sun, 31 May 2020 06:16:46 +0000 (06:16 +0000)]
efl_access_text: remove ptr usage in eo files
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11907
Woochanlee [Tue, 26 May 2020 06:36:54 +0000 (06:36 +0000)]
elm_gesture_layer: Arrange the logic for delete the target object in gesture cb.
When the user receives the callback of gesture callback, erases the target object, the gesture layer is deleted.
The memory is may broken and performing unnecessary operations during the logic.
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11838
Subhransu Mohanty [Wed, 27 May 2020 07:39:59 +0000 (07:39 +0000)]
eo: Fix memory leak in efl_key_wref_set() api.
Testcase:
elementary_test -to snapshot
Reviewed-by: Marcel Hollerbach <mail@marcel-hollerbach.de>
Differential Revision: https://phab.enlightenment.org/D11888
Hermet Park [Tue, 2 Jun 2020 00:59:51 +0000 (09:59 +0900)]
canvas svg: fix missing break.
JunsuChoi [Tue, 2 Jun 2020 03:20:37 +0000 (12:20 +0900)]
elm_hoversel: Add missing resize_job's free
Summary:
resize_job can be called after obj is deleted.
So add free to destructor.
Test Plan: N/A
Reviewers: Hermet, herb, kimcinoo, zmike
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11911
Change-Id: I5497ea6fa52198b30fee6d1f2283ea9a0141da8c
Kamil Konieczny [Mon, 25 May 2020 16:44:28 +0000 (18:44 +0200)]
elm_atspi_bridge: use efl_add_ref for obj without parent
Use efl_add_ref for object without parent, and unref
when obj is destructed.
This fixes error reported by EFL:
"E/EFL ( 956): eo<956> ../src/lib/eo/eo.c:1008
_efl_add_internal_start_do() Creation of 'Elm.Atspi.App.Object'
object at line 6882 in '../src/lib/elementary/elm_atspi_bridge.c'
is done without parent. This should use efl_add_ref."
Change-Id: Ib813c7b587c1e39d5335909037cec4664391dea3
Subhransu Mohanty [Mon, 1 Jun 2020 09:53:36 +0000 (18:53 +0900)]
svg/loader: Fix memory leak
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11908
Hosang Kim [Wed, 27 May 2020 11:15:15 +0000 (20:15 +0900)]
evas_object_main: add check seat grab count when touch list item is deleted.
Summary:
The touch event list is created by mouse down, mouse in or mouse move, etc.
But old list is deleted and new list is created in mouse down.
So it is necessary to distinguish list of mouse down and list of mouse move when deleting a list item.
Add checking seat's grab count to distinguish mouse down and mouse move.
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11890
Change-Id: Ic6066233a75885753d2abb19db802032bcaf0672
Signed-off-by: Hosang Kim <hosang12.kim@samsung.com>
Hosang Kim [Tue, 26 May 2020 11:13:49 +0000 (20:13 +0900)]
elm_conform: fix backward compatibility issue.
In Tizen 4.0, elm_conformant inheritted elm_layout, but elm_conformant
inheritted efl_ui_layout_base now. So elm_layout_file_set is not working
for elm_conformant.
elm_conformant inherits efl_file to fix backward compatibility.
Change-Id: If503eb709577902f13ab71e6e1dfb2c1f0507984
Prasoon Singh [Fri, 15 May 2020 15:55:38 +0000 (21:25 +0530)]
[edje] Free up the allocated array memory to avoid leak
Allocated array is only freed when it is inside if block but for
other case array is not freed.
Change-Id: I6b838f5908ee1c0f3f455dfa690c7ce30ee0471b
Signed-off-by: Prasoon Singh <prasoon.16@samsung.com>
Woochanlee [Fri, 29 May 2020 06:30:17 +0000 (15:30 +0900)]
efl_ui_clock: Ignore Second, Day format to backward compatibility.
The Second, Day fileds has been added, and that formats will generated.
It can breaks some use case old App. (auto format use case)
@tizen_fix
Change-Id: Id13b8445463fa26e9d4456051d5adf3e4cd9314a
Hermet Park [Fri, 29 May 2020 01:58:52 +0000 (10:58 +0900)]
edje edje_cc: remove beta tag for vector part.
vector part has been used since 2014, now it's widely used in many cases.
let's remove the beta tag, it's mature using in svg.
Change-Id: I1d0d795099c3cfabf70650657d0f5f9de5f09afd
Yeongjong Lee [Wed, 27 May 2020 06:17:45 +0000 (15:17 +0900)]
elm_bg: restore legacy bg sizing_eval to keep backward compatibility
In tizen 4.0, elm_bg size is always (0, 0). only internal image object had size
as file image size.
To keep legacy size policy, elm_bg override `elm_layout_sizing_eval` to
resize internal image and `efl_canvas_group_calculate` to prevent resize elm_bg.
Note that if you don't set any size to elm_bg, elm_bg size is (0, 0) by default
even though its child have size.
@tizen_fix
Change-Id: I54677a1024078ba1f157b8fd495e00dce713cce6
Woochanlee [Wed, 27 May 2020 10:44:59 +0000 (19:44 +0900)]
ecore_wl2: Add internal tag for Events.
Some of the Event should not be public
Change-Id: I603967d8a6a6d6e2f24829fea390b5757e0dddbc
SangHyeon Jade Lee [Tue, 26 May 2020 10:30:40 +0000 (19:30 +0900)]
genlist : support synchronized deletion in item select for compatibility safety
in tizen 4.0
all deletions are synchronized,
so user might write the code with expecting item is deleted by API call
immediately.
in tizen 5.5
we keep the object reference and defer the efl_del of item object in
special case that it has to alive,
and this action may broken backward compatibility for some application,
which is wrote wrong expectation,
but still we have to support them.
this code is fallback the tizen 4.0 deletion process.
item is referenced by efl_ref() while select callback is processing,
but internal item deletion is happen immediately,
and only the object itself is alive till we call efl_unref().
it is only activated by ignore_ref flag up,
for applying genlist tizen only.
Change-Id: I89e2645823bf603c084a2c2df7feb518798b2a36
Subhransu Mohanty [Wed, 27 May 2020 04:47:42 +0000 (13:47 +0900)]
elementary: Fix memory leak in elm_list
Change-Id: Ica703e4487b47f3724a229724941b4b8269db7fc
Subhransu Mohanty [Wed, 27 May 2020 04:08:35 +0000 (13:08 +0900)]
elementary: Fix memory leak in gengrid and genlist
Change-Id: Id21f707659f7eee031a9d367b82892eecb712dfe
Woochanlee [Wed, 27 May 2020 04:24:28 +0000 (13:24 +0900)]
ecore_wl2 : Add missing code when merge.
Upstream added this code in _xdg_toplevel_cb_close()
but we use _zxdg_toplevel_cb_close().
Change-Id: Ied2402f1baa44eb130e8e0d804feaa241ddde266
JunsuChoi [Mon, 1 Jun 2020 04:21:04 +0000 (13:21 +0900)]
Efl.Gfx.Vg.Value_Provider: Move internal function changed_flag to eo.
Summary:
changed_flag was declared in the private header for internal use(static_lib).
However, as the meson script has been changed, it can no longer be used.
changed_flag can display the property to which the changed
among the properties of value_provider.
there is no problem in providing it as an eo function, so move the function to eo.
Test Plan: N/A
Reviewers: Hermet, herb, kimcinoo
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Differential Revision: https://phab.enlightenment.org/D11894
Marcel Hollerbach [Mon, 1 Jun 2020 01:46:29 +0000 (10:46 +0900)]
elm_gen****: free item_content map
Summary:
this should be freeed. Both can be freed in group_del after the items
are deleted.
fix T8732
Reviewers: Hermet
Reviewed By: Hermet
Subscribers: cedric, #reviewers, #committers
Tags: #efl
Maniphest Tasks: T8732
Differential Revision: https://phab.enlightenment.org/D11882
Daniel Kolesa [Sun, 31 May 2020 05:26:52 +0000 (07:26 +0200)]
bindings/lua: simplify some eolian binding code
Daniel Kolesa [Sun, 31 May 2020 04:39:50 +0000 (06:39 +0200)]
bindings/lua: fixes for cffi-lua
Daniel Kolesa [Sun, 31 May 2020 01:39:49 +0000 (03:39 +0200)]
elua: fix object system on lua 5.2 onwards
This is a quick hacky fix, but it enables elua to work well with
lua 5.2+. Notably Eolian bindings work now.
Later this will be rewritten to use __gc directly on object
instances, with a fallback for newproxy for 5.1/luajit.
Daniel Kolesa [Sun, 31 May 2020 00:28:22 +0000 (02:28 +0200)]
bindings/lua: use new _fill api variants for expr values
The previous behavior was also invalid, since it was casting
a GC-managed pointer, which doesn't provide any guarantee that
it will stay valid. Fix that too, by using manually allocated
memory and assigning a finalizer at the end.
Daniel Kolesa [Sun, 31 May 2020 00:11:15 +0000 (02:11 +0200)]
eolian: add out-param variants of expr eval/value get funcs
This is for compatibility with bindings that can't express passing
unions by value (e.g. anything libffi based).
Daniel Kolesa [Sat, 30 May 2020 23:53:52 +0000 (01:53 +0200)]
bindings/lua: minor fixes to accommodate cffi-lua